Potential of Missing Life-Threatening Arrhythmias After Limiting the Use of Cardiac Telemetry

We previously revised our telemetry protocol by using the American Heart Association guidelines, producing a 43% decrease in telemetry initiation.^1,2 After determining that there was no increase in mortality, cardiac arrest, or activation of the rapid response team, we sought to ascertain the risk of missing life-threatening arrhythmias (LTAs) with reduced telemetry use. Life-threatening arrhythmias, such as ventricular tachyarrhythmias, are the primary rationale for using telemetry, and fear of missing them likely contributes to overuse. We studied the nature and clinical outcome of our telemetry alarms. We hypothesized that alarms representing LTAs are uncommon and that few alarms affect patient management.

In March 2013, we instituted a revision of non–intensive care unit telemetry that integrated the current American Heart Association guidelines² into our electronic ordering system. Predefined criteria (developed internally at our institution and in use for many years before our telemetry protocol revision) categorized telemetry alarms as emergency or nonemergency. Alarm events were communicated from a central monitoring department to patient care units via telephone, and a detailed log of alarms was maintained.

We selected 2 periods—before (October 19, 2012, to November 19, 2012) and after (May 22, 2013, to June 19, 2013) revision—to retrospectively review alarm logs. We tabulated the total number of alarms and studied, in detail, a random selection of alarm logs. All alarms designated as emergency in these subgroups were then evaluated with a detailed medical record review, arrhythmia analysis, and determination of whether any change in clinical management followed.

Emergency alarms were divided into 3 classes: potential LTAs, clinically important alarms, and alarms of questionable importance. Management changes included transferring to an intensive care unit, beginning use of a new medication, ordering a diagnostic study, or activating a rapid response or cardiac arrest team. The study was approved by the institutional review board at Christiana Care Health System. Informed consent was not required.

Statistical analysis was performed using the t test and χ² test.

Emergency alarms were infrequent (Table 1). All alarm logs for 1323 and 1322 randomly selected patients from the periods before and after revision, respectively, were examined in detail. The total number of alarms, examined in detail, was 4106 and 3094, respectively. There was only 1 potentially LTA alarm (0.01%) of the 7200 total alarms in these subgroups (Table 2). This patient had a self-terminated ventricular tachycardia that lasted 32 seconds. Thus, there was not a single LTA for which telemetry led to an immediate treatment during the study period. Of the 78 emergency alarms, 29 (37.2%) were classified as clinically important. However, only 14 (48.3%) of these 29 alarms led to a change in clinical management within 1 hour. Most of these alarms were for rapid atrial tachyarrhythmia. Telemetry length declined after the revision due to prespecified durations.

Even among the alarms designated as emergency, we found episodes of clinically important arrhythmias to be very infrequent, rarely leading to a change in patient management. Life-threatening arrhythmias were exceedingly rare, occurring in 1 of 2645 patients. Previous studies^3,4 have also found low rates of serious arrhythmias. In these studies, the incidence of LTA requiring immediate action was low, and there were few important changes in management or outcomes. For example, Schull et al⁵ reported that, of 8932 patients undergoing telemetry, only approximately 1 (0.02%) in 5000 were survivors of cardiac arrest in whom telemetry signaled the cardiac arrest.

Medicolegal concerns may contribute to telemetry overuse. We believe our system mitigates this risk as an evidence-based standard of care applied to clinical decision making through protocols. Thus, reducing unnecessary telemetry use is not likely to miss LTAs because of the very low incidence of true LTAs in contemporary telemetry monitoring settings. This finding should be reassuring to those considering the recommendation of the Choosing Wisely campaign to limit non–intensive care unit telemetry.⁵

Corresponding Author: Andrew Doorey, MD, Christiana Care Health System, 252 Chapman Rd, Ste 150, Newark, DE 19702.

Published Online: June 15, 2015. doi:10.1001/jamainternmed.2015.2387.

Author Contributions: Drs Kansara and Doorey had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Kansara, Dressler, Weiner, Kerzner, Weintraub, Doorey.

Acquisition, analysis, or interpretation of data: Kansara, Jackson, Dressler, Kerzner, Weintraub, Doorey.

Drafting of the manuscript: Jackson, Dressler, Weiner, Weintraub, Doorey.

Critical revision of the manuscript for important intellectual content: Kansara, Dressler, Kerzner, Weintraub, Doorey.

Statistical analysis: Jackson, Dressler, Weintraub.

Administrative, technical, or material support: Jackson, Dressler, Weintraub.

Study supervision: Kansara, Weiner, Kerzner, Weintraub, Doorey.

Conflict of Interest Disclosures: None reported.

Additional Contributions: Vasanth Chandrasekhar, BS (University of Delaware), contributed to the initial pilot study to determine the value of continuous cardiac telemetry at our institution, and Paul Kolm, PhD (Value Institute, Christiana Care Health System), provided statistical assistance. No financial compensation was provided.